Cite as: Stefanescu, D., 2020. Alternate Means of Digital Design Communication (PhD Thesis). UCL, London.

During the course of this thesis, a series of alternative approaches to digital design data communication were assessed. Taken together, they articulate a new technical platform for communication inside the AEC industry. Starting from a theoretical analysis of communication as both a social and technical phenomenon, three specific research questions were formulated that deal with distinct aspects of the communicative and collaborative act in the design disciplines, namely (1) representation, (2) classification, and (3) transaction.

From a methodological point of view, the research was embedded within a living laboratory context and, consequently, had an important applied component. This consisted of a new programmatic framework for digital design communication, named Speckle, which allowed us to analyse and assess (1), (2), and (3) against existing collaborative methodologies from the AEC industry. Speckle was developed throughout the research project as an open source, freely-available software platform that underwent continuous validation and testing both “in the wild”, from various industry stakeholders, as well as through three directed focus group meetings.

At a theoretical level, this research project argued that existing digital information exchange methodologies from the AEC industry enforce a technical model of communication on what is an essentially psycho-social and cognitive process, therefore leading to an unproductive design environment that restricts the emergence of shared understanding. By looking at (1), (2) and (3) from an integrated perspective of both technological affordances and social communication requirements, the research questions have expanded the tree aspects into the following research directions:

Data representation

what are the advantages and limitations of a lower-level, schema-abstract, composable object model over those presented by an ontologically complete higher-level standard?

The analyses in Chapter 4, Data Representation, have shown that ontological revision at a representational level—how design objects are defined—is a naturally occurring step throughout the design process. Sections 4.1 (Composable Data Structures) and 4.2 (Encoding Existing Ontologies), revealed that a lower-level, composable object model can simultaneously accommodate both ad-hoc, user-driven, data structures as well as programmatically match existing, pre-defined higher-level object models (such as IFC, or BHoM). Followingly, Section 4.4 (Managing Ontological Diversity) puts forward a methodology through which multiple object models can be supported in a digital communicative process without compromising on consistency and rigour.

The empirical analysis from Section 4.3 evidences that end-user driven ontological revision tends to minimise the complexity of the data structures involved, with a relatively large difference between the average tree depth of end-user created objects (1.97) and those coming from existing schemas (3.43). This translates in smaller object sizes (up to seven times) that, in turn, lead to a leaner digital data exchange volume.

Overall, the claim that, as an industry, AEC needs a singular and unique object model in order to consistently exchange digital design information is invalidated. Nevertheless, a major limitation is the issue of asymmetrical information codification and de-codification that occurs when communicants do not fully share the same knowledge context. This limitation, and its mitigation by enforcing one-directional information flows and object immutability, was addressed in Section 5.5 (Object Identity, Immutability & Data Deduplication).

Data classification

what are the advantages and limitations of an object-centred classification approach to digital design data as opposed to a file-centric collaboration methodology?

In Chapter 5, Data Classification, the analysis from section 5.2 (From Sharing Files to Curating Data) shows that, as opposed to a file-centric collaboration methodologies, an object-centric approach may impose ontological revision at the content level by requiring actors to negotiate what information they share and why. From the implementations and case studies presented in section 5.6 (Data History) and 5.7 (Applications), one finds that this leads to increased relevancy of data and a reduction in overall communication noise: instead of sharing information “in bulk”, any design data that is communicated must have a recipient and a direct use; furthermore the communicative network between stakeholders evolves organically.

The productivity of this curatorial approach to data communication was validated as well through the empirical analysis of data collected from Speckle. Sections 5.7.1 and 5.7.2 showed that end-users take advantage of the classification freedom offered by an object-centred approach by breaking down files in an average of two and half different groupings, and subsequently reassemble them in different “federated” models, containing an average of 2.78 different sources of data. In terms of actual storage size, the performance of the object-centred approach was empirically observed to be twice as high as a traditional file-based one (Section 5.7.3).

Nevertheless, the bottom-up evolution of “ad-hoc” federated models can prove to be too difficult to manage in a larger setting, and, as such, detrimental to the design process. This is because the sequentiality of information exchanges between actors, in a digital setting, is lost. As such, a key limitation is revealed as the lack of a clear communicative contract between participants.

Data transaction

what are the advantages and limitations of a centralised, file-based collaboration methodology as opposed to an object centric one with regards  to enabling the key requirements of a communicative contract, namely nextness and sequentiality?

Section 6.2 (Optimising Transactions: Differential Updates) shows that the transaction costs can be greatly reduced, in comparison with a file-centric collaboration model. Empirical observations from Section 6.2.2 show that they can provide a fivefold improvement, whereas their theoretical upper performance limit is not bounded, as they are directly proportional to the size of the change itself, rather than that of the whole model.

This is reflected in the observations from the case studies presented in Section 6.5: faster updates have the effect of allowing for the emergence of “nextness”, or immediacy, between information exchanges, which is a key quality of a communicative contract. As a result, the rate at which the representational revision identified in Chapter 4 and the content level revision from Chapter 5 proceed is increased, and, as such, act as a positive force for the emergence of shared understanding.

Section 6.3, Assembling the Network and 6.4 (Information Propagation) show that an object-centric approach can tie in data, author and source at any level of precision (from a whole model to an individual object) and thus allow for every stakeholder to be aware of what data they are depending on, from whom, and when it was last updated, thus making the sequentiality of a digital conversation tangible. The case studies from Section 6.5 show that this approach, while satisfactory for individual tasks, needs to be further studied so as to match the management and overview needs of a larger design process. Ongoing work in this direction was discussed in section 7.3 (Digital Transactions and Communication) of the Discussion.

Digital communication

Digital communication is a key infrastructural base on which the design process now operates. Within this research project, communication was contextualised as a transactional phenomenon, with both technical and social manifestations which reinforce each other. The contribution can be stated as an integrated technical and sociological reconceptualisation of communication in the digital design process that challenges the existing status quo of the AEC industry. By marrying contemporary technical affordances with a user- and industry-centred analysis, this project demonstrates that existing assumptions around the need for centralised high-level standards and workflows discourage meaningful dialogue to happen and exclude vital stakeholders from the design process. A flexible digital communication framework, by providing an inferential context for dialogue to happen amongst design stakeholders, allows for the emergence and evolution of the way information is defined and structured, enabling the creation of shared values and meaning. Broadly speaking, the low impact of emerging technologies on the overall productivity and efficiency of the AEC industries (Barbosa et al., 2017; Charef et al., 2019; Dainty et al., 2017; Hong et al., 2019) can be attributed to a confusion regarding the understanding of its communicative processes, and their subsequent distortion through inadequate technological implementations. Nevertheless, this research project concludes that digital technologies can embrace the diversity and richness of the design process, enhance the collaborative aspects of the industry, and, moreover, open an accessible and ethical pathway towards a digitally integrated built environment.